On Tue, April 12, 2016 2:11 pm, Sean Gallagher wrote:

Page 38 of this document states that the accuracy requirements are
measured relative to UTC(USNO), which would be US Naval Observatory.

http://www.gps.gov/technical/ps/2009-civil-monitoring-performance-specification.pdf

A google search for "how is USNO time transferred to GPS control" found
this document from the USNO web site:
http://www.usno.navy.mil/USNO/time/gps/usno-gps-time-transfer

--
Chris Caudle

Hi Sean,

(I)

The one official, legally traceable, source of precise time & frequency for the United States has always been NIST, based on their master clock(s) in Boulder Colorado. Many PhD's, over generations, from the Dept of Commerce, have brought you this amazing service. Yes, they excel in building ever more precise clocks; for greater USA commercial corroboration and prosperity.

The informal, but highly accurate, irresistible, and utterly ubiquitous source of precise time & frequency is GPS, which is a US military asset, based on master clock(s) in Washington DC, and also master clock(s) Schriever Air Force Base, Colorado. Many PhD's, over generations, from the Dept of Defense, have brought you this amazing service. Yes, they excel in building ever more precise PNT: Positioning, Navigation, and Timing systems; for USA defense and offense.

There are also other satellite time & navigation systems, with varying degrees of accuracy, trust, intent, commercial and military control, such as GLONASS (Russia), BeiDou (China), Galileo (EU), as well as Japan and India.

As time nuts in the early 21st century, we are showered with free precise time; we live in an unprecedented era where extreme precision flows through the air and down from the sky without regulation and for bargain prices.

(II)

Atomic clocks are good, but not perfect. All of these systems of atomic clocks regularly compare themselves with 350+ other atomic clocks around the world, under the coordination and authority of BIPM, based in Paris, France. So the same people that tell you how long a meter really is, or how massive a kilogram really is, or how much a volt really is, also have the job of telling us how long a second really is. That's the basis of precise time.

Note also the "C" in UTC is "coordinated", meaning both coordination amongst some astronomical clocks (earth, moon, sun) and also coordination amongst many terrestrial clocks. No one clock ever owns "the right time", not even spinning earth ball, so they are pooled together with a best-guess algorithm. So there is no truth, only coordination. That's why UTC is the standard; not perfect; but the standard.

(III)

To answer your question:

1. At any instant your GPS receiver simply gives you "the time" based on its plain cheap crystal oscillator. It will be close, but wrong.

2. But, that little crystal oscillator is gradually, numerically adjusted based on average coordinated reception of signals from up to a dozen fancy atomic GPS clocks in space. They will be very close, but wrong, at some level.

3. And, those atomic clocks in space are gradually, occasionally numerically adjusted to match the master clock at USNO. The master clock is closest, but even it will be wrong.

4. That GPS master clock, hydrogen maser MC2, at USNO is gradually, numerically adjusted based on the weighted mean of dozens of commercial cesium clocks, and three world-class, custom-designed, atomic rubidium fountain clocks at the USNO laboratory. It's a beauty, one of the modern wonders of the world; you should visit. USNO is the best in the world for PNT, but still not "perfect".

5. This whole set of USNO clocks is gradually, numerically adjusted based on the weighted mean of atomic clocks around the world. Any of those clocks will be close, but each still wrong if you look close enough. It's when you combine all of them that you get something better.

6. All of those clocks together are still a guess -- but since there are no better clocks in the universe to compare against -- this set of 350+ clocks on earth are called "right". And that is UTC. Everyone follows UTC. Someone has to tell us what the right meter is, the right kg, the right volt. BIPM tell us what the right time is.

So while it is true no one knows perfectly what time it is exactly now, through this multi-layered, world-wide, cooperative system -- everyone gets the benefit of everyone else's best clock. Everyone tries to be as close as possible, and yet everyone known they will be wrong at some level.

This is the mystery and allure of the world of precise time. Welcome to time nuts.

/tvb

----- Original Message -----
From: "Sean Gallagher" sean@wetstonetech.com
To: time-nuts@febo.com
Sent: Tuesday, April 12, 2016 12:11 PM
Subject: [time-nuts] Where does the source time for GPS come from?

Dear Sean,

The USNO is the main source of precise time in the US via the GPS system.
It has an ensemble of caesium clocks and hydrogen masers in Washington that
it uses as the reference to provide corrections for the clocks that are
onboard each of the GPS satellites. These corrections are broadcast
continuously via the GPS system.

What your boss is referring to is Coordinated Universal Time (UTC), the
international time standard. National labs which operate atomic clocks
compare their clocks using a variety of methods and send these to the BIPM,
which makes a weighted average of the all of the participating clocks (for
stability) and combines the ten or so primary frequency standard
evaluations (for accuracy) to finally get its best estimate of what one
second is. Each participating lab is then told what the difference between
its clock and UTC is.

USNO participates in UTC and keeps within tens of ns of UTC.

You can find the monthly reports in Circular T.

NIST also maintains an ensemble of atomic clocks and contributes to UTC.
However, in addition to commercial caesium beam standards and hydrogen
masers, they have several primary frequency standards, in particular,
caesium fountains.

Regards
Michael

On Wednesday, 13 April 2016, Sean Gallagher sean@wetstonetech.com wrote:

On 4/12/2016 3:16 PM, Michael Wouters wrote:

It might also be useful to introduce the concept of a "paper
clock", as opposed to a physical clock.  A paper clock
is where a lab has a physical clock that is compared to
UTC and they get an after the fact report that gives
them a correction factor to apply to their clock.  This
makes a paper clock.  So they can't tell you what time it
is now to a high degree of accuracy, but they can tell
you exactly what time it was last week.

Also note that there is a division of labor between
cesium clocks, which run continuously to keep time,
but may not be as accurate as "big" cesium standards
in terms of frequency.  However, the big standards
may not run 24/7 so they function as frequency standards
as opposed to true clocks.

There is also the whole issue of "time transfer" among
all these devices.  The gold standard is 2 way satellite.
But the everyday method is GPS itself.

Rick

On Wed, 13 Apr 2016 08:50:13 -0700
"Richard (Rick) Karlquist" richard@karlquist.com wrote:

At our visit at METAS i was told, that the Cs foutain is powerd
up once a week to measure the exact frequency of the H-Maser
and the three 5071's they have.

The gold standard is actually fibre. With that it's possible
to do sub-ns time transfer and <10^-16 frequency transfer (in a day).
But the longest fibre links are limited to a couple of 1000km.
Currently, there is a fibre link between Syrte (Paris) and PTB (Braunschweig)
over Strassbourg. It is planned to extend that to NPL (Middlesex)
and to INRIM (Turin). There is also a fibre from PTB to MPI-QP (Munich)
and it's likely that this will be integrated into the above network as well.
In east europe, there are also few fibre links, but these are considerably
shorter and have no interlinks yet (and no planned, as far as i am aware of).

		Attila Kinali

--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson